Clayton Kershaw and the thin line between one bad start and a catastrophe

Five days ago in the middle of the Dodgers' inexplicable sputter, Clayton Kershaw, of all people, failed. His outing -- 3 2/3 innings, nine baserunners and four runs allowed -- put the Dodgers in a hole they wouldn't climb out of. It was loss No. 7 in a team losing streak that has stretched to 11 entering Tuesday.

What happened? "It's not my job to diagnose it," Kershaw told reporters. "Just to see the results, and not like it."

OK, fine, but seriously -- what happened? What ever happens in a bad start? Kershaw starts again Tuesday night, and for the Dodgers -- not just Tuesday, not just this month, but potentially throughout October -- that question might be the key to the season.

There are three ways a good player can be bad: (A) He can actually perform as well as he always does but suffer to circumstances, luck, timing or a simply better opponent having his own great day; (B) he can be genuinely quite bad, but in the way that all of us have bad days and move on; or (C) he can be genuinely terrible, in a way that reflects an actual change in who he is, who he'll be going forward -- sore elbow, irreversible aging, etc.

The scary one is C, but in most cases, the answer to a bad day is A or B. For pitchers, I suspect, a far larger percentage of bad starts than we realize are A's. Philip Humber, the former major leaguer who once threw a perfect game, once told it to me like this: "To be honest, there's never much difference. The difference between a home run and a swing-and-a-miss is, what, an inch and a half? You can throw a great pitch, the guy makes a great swing. And if it's at a guy, it's an out." (In a blind test, most smart baseball fans couldn't really tell the difference between a pitch that leads to a hit and one that leads to an out.)

So we re-watched Kershaw's bad start against the Rockies from last Thursday, then we watched Kershaw's excellent start against the Rockies from June 24. We wanted to diagnose this bad start. What we saw is not great. In fact, most of what we saw showed up in the very first at-bat of the game, between Kershaw and Charlie Blackmon:

The first pitch was a fastball spiked in the dirt. The second was supposed to be outer half and sailed way outside. Two easy takes. The third was supposed to be low and away, and Kershaw missed up and over the plate. (It was fouled off.)

The fourth pitch was perfect. It was as fine a pitch as any Hall of Famer has ever delivered. Swung on and missed.

The fifth was supposed to be like the fourth, but it ended up less than half an inch from the exact center of the strike zone. (Foul.) The sixth, a slider, isn't on that plot, because it landed in front of the plate. The seventh was a fastball on the outer half. Charlie Blackmon singled.

A very generous reading of this at-bat -- and some others throughout the game -- might suggest we're talking about explanation A, the one that says Kershaw actually pitched well but got unlucky. On 2-1, he threw a perfect pitch, 92.1 mph right to Yasmani Grandal's target, right on Blackmon's hands, and right where a slider or a curve might have started if Blackmon had been looking for either of those pitches. He got a very good hitter to swing and miss on a 2-1 fastball. That's the good stuff. Kershaw threw a bunch of great pitches in this game. I promise.

And, though he allowed a hit to Blackmon, it was on a pitch arguably in a good spot, down and on the outer half, where Kershaw had tried to locate the first three fastballs of the game. The hit was just a ground ball that happened to be hit where nobody was standing. You could say the same about Mark Reynolds' weak single in the first, or Jonathan Lucroy's shallow sacrifice fly down the right-field line in the third.

But this explanation does not really hold up. The "arguably in a good spot" pitch Blackmon hit was supposed to be inside, and Kershaw missed all the way across the plate. The single Reynolds hit came on a pitch that was supposed to be away but ended up in the middle of the plate, belt-high. The Lucroy sacrifice fly came on a pitch that was supposed to be inside and missed over the plate:

These were bad pitches. In fact, we can say at the very least that we're talking about an Explanation B start: Kershaw was genuinely bad, relative to that June 24 outing. We tracked every pitch in each game and assigned them one of three colors based on location: Green pitches were great, yellow pitches missed but could reasonably have delivered good or bad outcomes, and red pitches missed badly. These are the sliders that bounced or ended up in the upper third of the zone or fastballs that missed not just by degree but by direction. (Pitch results were ignored; just intent and location.) Just let your eyes skim these:

June 24:

Sept. 7:

As you see, there are good and bad pitches in each start. This is crucial: In every start by every pitcher, there is at least one bad pitch that turns out well for him and one great pitch that turns out poorly. But if our unscientific red/green/yellow designations are to be believed, Kershaw made a lot of mistakes in this game -- twice as many as on June 24, with about half as many greens. Sometimes he got away with those bad pitches, as when he threw a terribly located fastball up and over the plate to Nolan Arenado in the fourth inning and Arenado fouled it off. Sometimes he didn't, as when he threw a terribly located fastball up and over the plate to Nolan Arenado in the first inning and Arenado homered:

On June 24, 74 percent of his sliders were strikes. Thursday, only 55 percent were. That's the key bullet point for Explanation B: Kershaw's slider was bad, so he had to throw more fastballs in more fastball counts, and a few of those fastballs got hit. Bad days happen. An optimist can believe Kershaw will walk out of the bullpen with better feel for the slider and presto! Back to normal. This is a compelling belief.

Explanation C, the scary one, hinges on the fastball, which showed both poor command and poor velocity. Kershaw's average fastball was 91.7 mph, the third-slowest start of his career. The only two starts with a slower recorded fastball: July 23 of this year, which Kershaw left early before going on the disabled list; and April 5, 2012, when Kershaw left his Opening Day start after three innings with the flu. (His excellent Sept. 1 start against the Padres, his first after coming off the disabled list, was not measured by PITCHf/x because of hardware problems.)

Low velocity starts happen, too, and Kershaw -- just off the disabled list and with a 10-game lead in the NL West -- might have had reasons to take it easy. But there is circumstantial evidence he really was laboring to throw even as hard as he did. Remember the fifth pitch of the game, the fastball down the middle to Charlie Blackmon? Kershaw delivered that pitch with an audible grunt, and it was his hardest of the night, at 93.14 mph.

It was located terribly, which is true of almost all of Kershaw's hardest pitches. Of the 10 fastest pitches he threw in the start -- all of them 92.3 mph or harder -- seven of them were "red" on our log, for being poorly located. Of the other 35 fastballs he threw -- all under 92.3 mph -- only seven more were red. That suggests, if doesn't prove, that he was straining to get even to 93. He threw one fastball -- on 1-0 to Trevor Story -- under 90 mph. He got one swinging strike on the pitch all night.

So here's what we know: Like all pitchers in all bad starts, some things went against Kershaw that didn't have to. However, Kershaw's slider control was off, and his fastball command gave the Rockies far more hittable pitches, or easy takes, than in a typical Kershaw start. That wouldn't be too troubling on its own, except Kershaw's velocity was down and his poor command might have been caused by his trying to make up for that lost velocity. Perhaps not coincidentally, he recently missed an extended period with back pain, which caused his only start over the past five years with similar low velocity. What we don't know: Whether we've just described one bad day or a bad month.

The Dodgers are in a terrible place right now, but the benefit of having a double-digit lead in one's division -- as they did at the start of this slump -- is that a long losing streak doesn't change much. The Dodgers' chances of winning the World Series are barely affected by whether they end up with 100 wins or 110. What would affect those chances -- greatly -- is seeing Explanation C bad starts from any of their top four starters. Alex Wood's velocity is already down, and his ERA and FIP are climbing. If Kershaw's velocity follows, the Dodgers would be, well, not exactly underdogs in October, but long shots like everybody else.

So we'll be watching Kershaw closely Tuesday night to see whether he pinpoints a 94 mph fastball in the first inning, whether at some point he touches 95 as he often does, whether he sits comfortably at 93 without missing on the wrong side of the plate. It's not his job to prove this fear wrong. Just to see the results, and like it.